Our Blog

Lauren Golembiewski

Episode 54: Conversational Interface & Chatbot Technology: Everything You Need to Know, with Lauren Golembiewski

Lauren Golembiewski is the CEO and co-founder of Voxable, a conversational interface agency. She’s lucky enough to run Voxable with her partner in life, Matthew Buck. Lauren and Matthew founded Voxable to help clients more effectively communicate with their customers through conversational interface design and development. They were originally inspired to create Voxable when they connected their home to a voice-first device and felt like the wizards and space captains they saw on TV.

Lauren’s background is in product design and user experience which is driven by her fascination with the way things work, especially the human mind. Her experience working with startups and leading design strategy for tech companies enables her to dive into design problems and help clients understand the best path forward. She regularly speaks about conversational design and advocates for a greater creative effort to be invested in this new technology.

What you’ll learn about in this episode:

  • Chatbots, Alexa skills, actions on Google, and more: what Voxable builds as a conversational interface agency
  • How these conversational interfaces help humans and machines better understand each other
  • How these conversational interfaces take human questions or commands (either verbal or written) and turn them into a machine response so that the machine does the work for the human and examples of each of these in practice on the various platforms these bots can be built into
  • The research that needs to be done before building a conversational interface to understand how a business responds to its customers now and assess ways to improve communication with a bot
  • Building a bot that serves your customers in the way they want to be served (ex: sends them a helpful blog automatically)
  • The work that needs to be done to write all of the lifelike responses that a bot would provide a human during an interaction
  • Coming up with every scenario that a user would interact with a bot and coming up with paths and responses for users to reach their desired outcomes
  • How bots can remember user answers to better steer them towards their prefered goal
  • The importance of having a style guide for how the bots communicate with users
  • How Voxable’s clients test the system throughout different points in the developmental process
  • Tweaking the CMS that a bot is built on to push out whatever message is important at a specific time
  • What happens if a bot doesn’t understand a user or can’t help with what they ask for
  • How users interact with your bot when it’s a persona that you’ve built well for your brand
  • How to measure whether a bot is working or not

Ways to contact Lauren:

Podcast eBooks:

The Power of Two

Episodes 1, 2 and 3 collide to bring you summary of lessons learned and systems created around Vision and Key Initiatives that help drive success to companies and businesses.

The Transition to Automation

In Episode 25, Vera talks with Heidi Rasmussen, CEO and Co-Founder of one of Inc 5000’s fastest growing companies in America – freshbenies. This eBook highlights part of the conversation to bring out the best lesson in automation and on-boarding for startups.

Using IT Strategically

In Episode 29, Vera talks with Tom Grooms, Vice President, Information Technology, and Chief Information Officer for CF Industries. This eBook is your guide for seeing IT as more than just a faster way to do your accounting.

The ZFactor Methodology

In Episode 35, Vera talks with Cindy Goldsberry, founder and partner of ZFactor Group. This eBook shows you how to take your business from vendor to value creator.

Transcript:

Welcome to System Execution, the strategy and system behind today’s successful companies. Systems can make or break your company. But here we’ll solve your physical, technological, and psychological systems issues by connecting you with experts that have succeeded in overcoming those challenges in their own business. And providing your the guidelines and tools you need to implement those same strategies for immediate results. Now, here’s your host, Vera Fischer.

Vera Fischer: Today’s episode is sponsored by 97 Degrees West, the brand marketing agency located in Austin, Texas. 97 Degrees West serves regional and national companies in the healthcare, finance, energy, and manufacturing industries. 97 Degrees West believes that an integrated approach to marketing that involves traditional and digital strategies that fit your customers’ buying journey yields the greatest impact on your bottom line. Go to www.97DWest.com to learn more.

Welcome to System Execution, a podcast devoted to using processes and systems to drive to a better outcome for your business. I’m Vera Fischer, your host. Many of your know that business success relies on systems. Systems can be physical, such as a warehouse or a factory. Or technological, think software. While others are psychological systems, such as checklist, org charts, or your daily hot list.

I’m really excited today to announce my guest, her name is Lauren Golembiewski, she is the co-founder of Voxable, a conversational interface agency. And we’re going to get to what that actually means here in a few minutes.

But Lauren is really lucky enough to run Voxable with her partner in life, Matthew Buck. Lauren and Matthew founded Voxable to help clients more effectively communicate with their customers through conversation interface design and development. She regularly speaks about conversational design and advocates for greater creative effort to be invested in new technologies. Welcome to System Execution, Lauren.

Lauren G: Thanks Vera, it’s great to be here.

 

More on Lauren’s Background

Vera Fischer: Well Lauren, it’s really great to have you on the show. And before we dive into your conversational design process, if you don’t mind, please expand on who you are, what you’re about, how you started Voxable, as I know my listeners really love to hear that.

Lauren G: Absolutely. You’re exactly right, we are a conversational interface agency, which means that we build conversational interfaces for other companies and products. And that includes chat bots, Alexa skills, actions on Google, and a conversational interface that might be embedded into a web application or mobile application. And really our mission is help humans and machines better understand each other to foster better communication in business.

And we originally got really excited about this space when we connected our home automation devices, our internet of things smart devices, to voice first devices like the Amazon Alexa and the Google Home. And those were the most popular ones, but we actually had a predecessor that we could basically hack and program new experiences for ourselves. And we really found that to be magical and amazing.

And then what we noticed is that there is a paradigm shift happening in using some of these devices, where human to computer interaction is actually doing more for the user. Doing more work that humans would typically have to do. So standing up and walking over to a light switch, now I can simply ask my home to turn on its light switch. And conversational interfaces are a really big part of that shift. So that’s really the value that we saw. And we knew that businesses could really benefit from that value as well.

So when we make an interface conversational, we’re saying to the user, “Hey, you can approach us with the thing that you need, and just say it to our app, website, or bot, and our software intelligence will go and find or do that thing for you.” A user no longer has to hunt through an app or a website. Their goal can be immediately met just by asking or typing their question into a chat. Quite literally, machines are being taught to understand and speak our language. And we are helping businesses speak their customers’ language.

Vera Fischer: So Lauren, that’s really cool. And I know that for some of our listeners, they are very fluent in the language that you’re speaking right now. But let’s give our other listeners a little bit, a few minutes on kind of educating us on exactly what a chat bot is, or what these interfaces are, in English terms, if you will.

Lauren G: Yeah, absolutely. So there’s two modes of interaction in a conversational interface. And that is either via voice, or talking to the interface, or via text, chatting with the interface in a messenger. And so the applications that we build actually are inside of a messaging app or used through a voice first device like Alexa. So the most popular chat app that we build our applications into is Facebook Messenger. Facebook Messenger has built a platform for businesses to leverage their messaging platforms to reach out to their customers. It has a very robust set of tools. And that’s actually what the bot is. It’s an application that lives inside that messaging channel that automates communication between customers and business, or the audience and businesses.

In the voice space these interfaces are attached to devices like the Google Home or the Amazon Alexa. And there’s a lot of things that come out of the box with those devices, like setting an alarm or a timer, or asking about the weather. But we build other applications and we help businesses develop strategy to make their content readable by these voice devices. As well as develop applications that allow their users to control aspects of that particular application via their voice.

The voice space is really interesting. Not only is it available in voice first devices but I can be embedded directly into a website or a mobile app. So there’s so many opportunities to actually make the voice space much richer and connected to many more devices than we typically use today.

 

The Conversational Interface Design Process

Vera Fischer: All right. Well, thank you so much for that education, I loved that. So let’s get started and let’s go through this conversational design process. So take us through this process. What would be the first thing that you guys would get started with from, if a customer was coming to you and said, “Hey, I’m really interested in this design process. How do I get started?”

Lauren G: Yeah, absolutely. There’s really three large sections of our process that encompass strategy, design, and development. And within strategy we really focus on performing user research, understanding a company’s audience, a business’s audience, and analyzing their current modes of communication to their users today. So how do they currently communicate via social networks and email? What does their existing communication look like? And where can they improve? And where are the opportunities to reach more customers through something like a Facebook Messenger chat bot or an Alexa device?

And so we dive into research to really understand those customers and the business opportunities within them. A lot of our research is around jobs to be done, which is a framework that positions your tool or your piece of communication as something a user or a member of your audience would hire your particular application to do. So a lot of times folks hire certain blogs to give them information about current events or current info within their industry. And so really thinking about what users are hiring your company to do when they’re engaging with you. And finding opportunities to better serve those jobs and those needs of the user.

Vera Fischer: So Lauren, it’s something important that I want to point out to everyone is a lot of folks think that somehow these conversations are just, they just appear. So if you’re out there and you’re on a website and you want to do some type of a chat, in some cases those are pre-programmed answers to whatever questions you’re asking. But someone actually has to create the responses to ensure that it feels like a human being on the other side. Would that be accurate?

Lauren G: Yeah, absolutely. And that is definitely why we would like to understand current customer communications. So that we can really speak their language and craft the right conversation, the right conversational messages to serve those customers.

Vera Fischer: And when you’re developing those conversations, so you’ve gone through that first step of really understanding the customer for that particular client, maybe the different scenarios in which they would be speaking through that messenger. Is that right?

Lauren G: Yeah, exactly. We really focus on, and as I mentioned, jobs to be done really focuses on the situation, the motivation, and the desired outcome for each particular job or goal that a user has. And so we really want to create a lot of, or understand a lot of the scenarios in which users approach businesses.

Vera Fischer: Okay, so once you’ve gone through that, so how do you, you must map out lots of different conversations. How many do you typically do?

Lauren G: Oh, it gets pretty big. We start with the most important, really, what are the ones that are going to have the highest value for a business? But we do map out large, that’s exactly what our design process looks like, once we understand the goal then we need to model a conversation that gets that customer to a goal. And then we also have to think about all the different paths that that conversation could branch off into.

So, where do customers actually want to go? And how do you guide them to a goal that they have? And so we map, we make large maps. It can comprise of anywhere between 50 to a couple hundred different threaded threads of conversation that we’re actually mapping out and looking at and thinking, “Okay, does this make sense for how we’re serving our customer?” And often we even act out some of those conversations and we’ll have someone read off scripts and walk through a human to human conversation in our mapping.

 

How Long this Process Takes

Vera Fischer: Oh wow, that’s cool. I didn’t think about that part. So Lauren, how long do you allot in your process as far as time goes, to develop and act out those scenarios?

Lauren G: It usually requires about a month to six weeks to really map out all of the different scenarios and review them with our clients and make sure that we’re capturing the right variations and the right contexts. So contexts can switch very often in a conversation. We also try to make sure that we’re remembering things about the user, and that really feeds into that contextual awareness. So if the user says they like a color, a certain color of a product, then we might try to remember that and intelligently give them better answers and direct the conversation in a different way. That’s a really simple example, but there’s all types of things that we’re trying to map out and account for when we’re creating these conversations. And we really iterate, like I said, we act them out, we make sure that the tone feels good. We write style guides for the conversational brand voice tone, and the possible persona character that might be embodying that conversation. So that whole process is really the core chunk of design work that we’re doing. And we hand those assets to an engineering team or someone who may be implementing that conversational interface.

Vera Fischer: Okay. All right, so you’ve done that, and assuming that it’s all signed off and approved, et cetera, by both teams, the client and Voxable, then what happens?

Lauren G: Then it goes into engineering or implementation. So, there’s some existing tools that we can use to build a conversational, simpler conversational interfaces. The more complex ones, that may be saving more complex user profile information, or maybe pulling from the user profile information you already have as part of your business intelligence, those more complex applications are developed by my partner, his team, he’s the CTO, and they build the conversational interfaces. And that can take maybe another four to six weeks as well. And there we’re trying to quickly get small pieces of those conversation threads built and then actually using them as the developers build them out, to make sure that in the channel that we’re using, whether it’s voice or those messaging channels, that it feels good and it’s working the right way and we have the right affordances accounted for within those different platforms.

Vera Fischer: So, how does the client, how does your client know that they’re working?

Lauren G: So, they can use them directly. And we actually give them previews of our current process and test them. So there’s some different tests, that we may prototype a specific interaction that we think might be a little bit more difficult before we actually build it, so that our client can get a preview before we invest our developers’ time into it, or a significant portion of our developers’ time into it. And then throughout the development process we release the most important pieces of the app, the highest level, the ones that we think are going to have the biggest impact. And we have an agile process, so we try to break it out into small enough chunks that we can deliver working pieces of the conversational interface to our clients so that they can actually experience it for themselves.

Vera Fischer: So, whenever you’re letting, or you’ve turned this over and your clients are using it, what if they’re wanting to run some type of a special, and that’s part of that mapping? Is it easy to go in and change it up based on, “I’m a retailer and it’s Christmas or it’s Valentine’s or it’s Memorial Day.” Is that easy to do, or do they need to come back to you to have you tweak those maps?

Lauren G: So it totally depends on the specific implementation. We often either connect to existing content management systems or build a purpose-built content management system for the conversational UI, so that clients do have the ability to change the content that exists, as well as push out new content. So that’s kind of the broadcast or push messaging that is available in chat, not totally available in voice.

But in the chat channel you can actually kind of send kind of broadcast messages. And those will send a push message to your audience that has subscribed to your bot, and they could receive updates about your company. And so we could connect a blog so that your new blog posts go out to the customers that it makes sense to push those blog posts out to. If you have particular segments, you can segment your broadcast messages to the right folks in your audience. And so these content management systems help companies manage all of the bot’s content.

A lot of companies want to start small and they want to just see the chat bot in place and put it out there into the market and see how it behaves. And so they may not invest into a content management system right away. And in that case, they are coming back and having us change tweaks here and there, and add maybe new questions or new goals for the customers one bit at a time. But really, after we’ve built a foundational conversation, making those little updates is fairly easy and not really super time intensive.

 

Mistakes in the Process

Vera Fischer: Okay, all right. Well that sounds cool. So let me ask you this, whenever you are implementing this chatbot, are there any things, or anything that can happen that can kind of mess it up or have an unintended consequence that no one foresaw? I mean, just something that our listeners could be aware of.

Lauren G: Yeah, absolutely. The conversation is inherently unpredictable, even from a human to human conversation, you never really know where it can take you. And that’s the exciting part of conversation. And I think that’s why it’s really exciting for users, that they can kind of choose their own adventure. But again, when users are testing the edges or trying to figure out exactly what your conversational interface can do for them, they’re going to ask some unpredictable things. So the way we really like to account for that is, again, the research will help us really know what our customers want and need, so even if they’re trying to diverge or do something that’s unrelated, that we can kind of say, “Hey, this isn’t something that we do, so here’s how we can help you.”

The other cases are the bot might not understand the user. It might take them to a place that they didn’t want to be, because we didn’t account for all of the unpredictability of conversation. So really we spend a lot of time on trying to understand where errors occur in conversation, and try to gracefully get those users back on track.

And then we also build in a lot of, this is where branding is really important when we talk about conversational interfaces, is we try to build a back story around the conversational interface that a company is releasing to their users. And the reason that’s important is because users project their own human emotions onto these interfaces, more so than any other interface that really exists today. So they start to think about these interfaces that are talking back to them, or sending texts back to them, as kind of a personified character that is evoking emotions within them, and they do things like say please and thank you. And things that you wouldn’t necessarily think are the same social rules applicable to these machines that you’re discussing with, that you’re having conversations with.

So it’s back to branding, really honing in on how your voice, tone, and persona speaks through these conversations helps guide the users to what’s available. And can kind of delight them when they ask them something about, you know, “Who created you?” or, “Where were you born?” or, “How are you today?” Those kind of small talk ways that we connect as humans, but that’s also the way humans are starting to connect with machines.

Vera Fischer: Well, and you know, that is a very important point that you’re making, is a lot of business owners, business leaders, may not have thought about their brand in terms of a persona, in the sense of, “How would my brand speak to someone in a one to one conversation?” And that’s really where it’s going right now, and that’s what this is. Would you agree?

Lauren G: Yes, absolutely. And it doesn’t necessarily have to be a mascot that you create, like a totally different character. Even though that’s very popular and we have seen success with that resonating with users. Some brands may choose to be a little bit more reserved and still continue their core brand expression, but just translate that out to a conversational interface.

 

How to Measure its Effectiveness

Vera Fischer: Gotcha. All right. So, now when you’ve got this system all up and running, the chat bot or the conversations, how do you measure the effectiveness of this particular system?

Lauren G: So we really like to perform internal testing as well as user testing to gauge its effectiveness before we really put it out there in the world. And that really just helps us to make sure that internally that experience is working and we think it’s doing the things that we set out to design. And then we might pilot it with a small group of folks who are external audiences. It really all depends on how important or how far out on a limb we’re going with deploying this interface, as to how much we really take in rigorous testing before we release. But we do some of that. And then when we release the bot we also make sure that we have good analytics, we’re focusing on the right metrics. And so that goes back to the goals that we defined originally in the project. And understanding, “Okay, what are the signals that we’re going to see in these conversations that are really telling us that we’ve done something that’s resonating with our audience?”

And so there’s easy goals like, “People are chatting a lot, they go back and forth, they’re engaging with our content, they’re opening up the blog post that we sent them through that bot. And maybe they’re sharing it with their friends.” And then there’s the more difficult ones like, “Oh, this person seems to be engaging with us in other places, they came to our bot, they had some conversation and they never returned. So how do we recapture some of those folks, or how do we understand how we’re not serving them?” And so those start to become a little bit more difficult and it kind of starts a whole, it can kick off another research endeavor of just understanding how people respond to the interface that we’ve just opened up to the world.

Vera Fischer: Interesting. Well, I know it’s a new frontier, and I know that you and your co-founder are somewhat at the forefront of this frontier, so I’m very excited about it and I know that it’s, those who really take it seriously and do the implementation well I think will get some really good results out of it.

Lauren G: Yeah, absolutely. I totally agree. And the market is still new, but we’re seeing a lot of really positive metrics that tell us that this is the way that folks want to engage. Especially younger audiences who are not really on email as much, who are messaging every day with their friends, who want quick and personal interactions from businesses. They want to be heard and they want someone to respond to them as soon as they’ve messaged a business maybe with an issue, or maybe just with, “Hey, I want more content. I want to see more of what you’ve got.” And that is really what conversational interfaces provide for those businesses that do move on them more quickly.

 

Next Challenges for Lauren and Voxable

Vera Fischer: Well Lauren, I’m just loving all the information that you’ve shared with us around this conversational interface design process. You’ve given us some great takeaways. So before we wrap up our discussion, what’s your next challenge with Voxable?

Lauren G: So we’re really looking at how the voice space is going to evolve a bit more and really become part of our everyday lives as we connect more and more devices. So we’re really excited about the opportunities within voice and are starting to build a little bit more internal demos and try some new things.

On the chatbot side, we actually have an internal product that we’ll be releasing early next year, that is a really fun way to shop and find gifts for folks. And so we’re kind of putting out our own internal bot because we have seen so many, we see so many opportunities out there and so many use cases that are really valuable, and we don’t always get to work on them with our clients. We don’t always have clients who want to build the things that we see are huge opportunities. And so we’re like, “Well, we can build them ourselves.” So that’s what we’re starting to do more and more, as we take on client work and trying to carve away some time to build some of these internal bots that Voxable owns. So we’re going to be releasing some fun things early next year that I think will delight folks.

 

Final Advice from Lauren

Vera Fischer: Well Lauren, you’ve shown us that processes are needed to get the work done, and you’ve provided a few of the nuances that our listeners need to hear regarding the execution of a successful system. Before we go, let’s close out today’s discussion with any final advice you want to share, anything that we may have missed. And then tell us the best way we can connect with you.

Lauren G: Yeah, so we really love and use systems throughout our entire process in our company and we really believe that all interfaces are built on top of great systems and processes. And so I really love that we get to talk about, that you’ve given me the opportunity to talk about this with folks as it relates to our particular business.

And the way that you can get in touch is you can visit our website at voxable.io, our Twitter handle is also @voxable. If you want to reach out to me personally, I am @laurengolem on Twitter. And we would love to hear feedback or talk to folks who are interested in learning more about the space.

Vera Fischer: Well, System Execution fans, no matter how many notes you took or how often you re-listen to this episode, remember, every successful business uses systems to drive to a better outcome. Lauren, it’s been great to have you on the show and thank you for sharing your insight with System Execution listeners today.

Lauren G: Thank you so much, Vera.

We hope you found this episode of System Execution enlightening. For free examples, case studies, e-books, and more, be sure to visit systemexecution.com/resources. Contribute to the conversation by reaching out to Vera directly on email at vera@systemexecution.com. Until our next episode, thank you for the privilege of your time.

Comments

comments

Tags:

This is a unique website which will require a more modern browser to work! Please upgrade today!